what models/analysis were you thinking of carry on if you have already read some literature on the topic, please list the references.
Implied volatility probability distributions (IVPD) are a way of understanding what the market is expecting regarding the future price of an asset (both stocks, indexes and ETFs) based on the price of options available. showing the whole distribution of probabilities and the associated strike prices, essentially what is the most probable price for the future.
The study of implied volatility distribution functions (IVDFs) forms a crucial bridge between derivatives pricing theory and the empirical analysis of market expectations and risk preferences. Formally, the IVDF is based on the Risk-Neutral Density (RND), representing the market’s expected distribution of the underlying asset price at a future expiration date, calculated under the risk-neutral measure \(\mathbb Q\).
This to tackle both descriptive analysis and create a framework with practical forecasting applications.
Inuputs: Price \(P\): The price of the option incorporates expectations of the market operators on the future volatilty of the asset.
Implicite Volatility: Stemming from the price of each option, we can calcolate the “implicite volatility”. It is the volatility that, when inserited in a model like Black-Scholes, would give you back the market price.
Volatility Smile/Skew: after calculating the single Implied volatility with options for options with different strike prices \(K\) with the same maturity \(T\), it’s not creating a smooth line like what a risk neutral probabilty assumed by B&S, it takes on the shape of a smile or a skew, indicating that the market is expecting the probability of different moviments for extreme prices.
Probability distribution: Costruzione della Distribuzione: Utilizzando queste volatilità implicite per l’intera catena di opzioni, si possono costruire delle curve che rappresentano la distribuzione di probabilità neutrale al rischio del prezzo futuro dell’asset. In pratica, queste curve ti dicono quali prezzi futuri sono considerati più probabili dal mercato e quali meno. it could also be bi-modal, the market is either hoping for an increase or a decrease leaving almost no space for the near strike events.
The most important feature of this distribution is where the distribution is centered in term of the strike value, symmetric compared to the current price, is the market expecting a price rise or a fall. If instead the distribution is asymmetrical (skewed) indica una maggiore probabilità di movimenti in una direzione (es. più probabile un calo che un rialzo).
“Fat Tails”: Se le code della distribuzione sono più “pesanti” del normale, significa che il mercato attribuisce una probabilità maggiore a eventi estremi (grandi rialzi o grandi crolli) rispetto a quanto previsto da un modello standard.
3D Changes in time: once you have one distribution for one maturity date you can look at how these distributions change over time 3d view
4D Changes in time: changes before and after big important events (earnings call or the announcment of the launch of a new product), we can understand the how the expectations of the market are modified when presented with new informations.
An uncertain event like the covid pandemic coudl flatten the distribution, indicating more uncertainty,
Move the peak of the distribution, indicating a change in the direction of the expected price.
assuming that there are no market shenanigans
import yfinance as yf
import numpy as np
import pandas as pd
from scipy.stats import norm
from scipy.optimize import brentq
import plotly.graph_objects as go
# import plotly as plo
# plo.renderers.default = "notebook_connected" # or "iframe"
from plotly.subplots import make_subplots
from scipy.stats import norm
from datetime import datetime
import matplotlib.pyplot as pltThe core concept is based on Breeden and Litzenberger (1978), who showed that the second derivative of a European call price with respect to strike gives the risk-neutral PDF of the underlying asset price at maturity: \[ f(K)=e^{rT} ⋅ \frac{∂^2C(K)}{∂K^2} \]
def black_scholes_price(S, K, T, r, sigma, option_type="call"):
"""Black-Scholes price for call/put"""
d1 = (np.log(S / K) + (r + 0.5 * sigma**2) * T) / (sigma * np.sqrt(T))
d2 = d1 - sigma * np.sqrt(T)
if option_type == "call":
return S * norm.cdf(d1) - K * np.exp(-r*T) * norm.cdf(d2)
else: # put
return K * np.exp(-r*T) * norm.cdf(-d2) - S * norm.cdf(-d1)
def implied_volatility(market_price, S, K, T, r, option_type="call"):
"""Solve for IV using Brent root finder"""
try:
f = lambda sigma: black_scholes_price(S, K, T, r, sigma, option_type) - market_price
return brentq(f, 1e-6, 5.0, maxiter=500)
except:
return np.nan
def implied_pdf(S, K, T, r, sigma):
"""Breeden-Litzenberger: risk-neutral density via BS second derivative wrt strike"""
d1 = (np.log(S / K) + (r + 0.5 * sigma**2) * T) / (sigma * np.sqrt(T))
d2 = d1 - sigma * np.sqrt(T)
pdf = (np.exp(-r*T) * norm.pdf(d2)) / (K * sigma * np.sqrt(T))
return pdfdef calculate_probability_distribution(strike, spot, iv, days_to_exp, risk_free_rate = 0.05):
"""
Calculate the implied probability distribution from Black-Scholes
Returns price range and probability density
"""
T = days_to_exp / 365.0
if T <= 0 or iv <= 0:
return None, None
# Generate price range (±4 standard deviations)
std_dev = spot * iv * np.sqrt(T)
price_range = np.linspace(max(spot - 4*std_dev, 0.01), spot + 4*std_dev, 200)
# Calculate log-normal distribution parameters
mu = np.log(spot) + (risk_free_rate - 0.5 * iv**2) * T
sigma = iv * np.sqrt(T)
# Probability density function
pdf = (1 / (price_range * sigma * np.sqrt(2 * np.pi))) * \
np.exp(-0.5 * ((np.log(price_range) - mu) / sigma)**2)
return price_range, pdf
def get_option_chain_data(ticker_symbol):
"""
Fetch complete option chain data for all expirations
"""
ticker = yf.Ticker(ticker_symbol)
current_price = ticker.history(period='1d')['Close'].iloc[-1]
expirations = ticker.options
all_options = []
for exp_date in expirations:
opt_chain = ticker.option_chain(exp_date)
# Process calls
calls = opt_chain.calls.copy()
calls['type'] = 'call'
calls['expiration'] = exp_date
# Process puts
puts = opt_chain.puts.copy()
puts['type'] = 'put'
puts['expiration'] = exp_date
all_options.append(pd.concat([calls, puts]))
options_df = pd.concat(all_options, ignore_index=True)
# Calculate days to expiration
options_df['expiration'] = pd.to_datetime(options_df['expiration'])
options_df['days_to_exp'] = (options_df['expiration'] - datetime.now()).dt.days
# Filter out options with missing IV
options_df = options_df[options_df['impliedVolatility'].notna()]
options_df = options_df[options_df['impliedVolatility'] > 0]
return options_df, current_price
def create_3d_ipdf_surface(options_df, current_price):
"""
Create 3D surface plot showing how IPDF changes across strikes and expirations
"""
# Group by expiration
expirations = sorted(options_df['expiration'].unique())
# Prepare data for surface plot
price_ranges = []
pdfs_matrix = []
exp_labels = []
for exp in expirations:
exp_data = options_df[options_df['expiration'] == exp]
days_to_exp = exp_data['days_to_exp'].iloc[0]
# Use ATM option IV as reference
atm_option = exp_data.iloc[(exp_data['strike'] - current_price).abs().argsort()[:1]]
if len(atm_option) > 0:
iv = atm_option['impliedVolatility'].iloc[0]
price_range, pdf = calculate_probability_distribution(
current_price, current_price, iv, days_to_exp
)
if price_range is not None:
price_ranges.append(price_range)
pdfs_matrix.append(pdf)
exp_labels.append(f"{exp.strftime('%Y-%m-%d')} ({days_to_exp}d)")
# Create meshgrid for 3D plot
if len(price_ranges) > 0:
fig = go.Figure(data=[go.Surface(
x=price_ranges[0],
y=list(range(len(exp_labels))),
z=pdfs_matrix,
colorscale='Viridis',
opacity=0.8,
name='Probability Density'
)])
fig.update_layout(
title=f'3D Implied Probability Distribution - Full Option Chain looking at {TICKER}',
scene=dict(
xaxis_title='Stock Price',
yaxis=dict(
title='Expiration',
ticktext=exp_labels,
tickvals=list(range(len(exp_labels)))
),
zaxis_title='Probability Density',
camera=dict(eye=dict(x=1.5, y=1.5, z=1.3))
),
height=700
)
return fig
return None
def create_ipdf_overlay(options_df, current_price):
"""
Create 2D plot with multiple IPDFs overlaid with alpha transparency
"""
fig, ax = plt.subplots(figsize=(12, 7))
expirations = sorted(options_df['expiration'].unique())
colors = plt.cm.tab10(np.linspace(0, 1, 10))
for idx, exp in enumerate(expirations[:8]):
exp_data = options_df[options_df['expiration'] == exp]
days_to_exp = exp_data['days_to_exp'].iloc[0]
# Use ATM option IV
atm_option = exp_data.iloc[(exp_data['strike'] - current_price).abs().argsort()[:1]]
if len(atm_option) > 0:
iv = atm_option['impliedVolatility'].iloc[0]
price_range, pdf = calculate_probability_distribution(
current_price, current_price, iv, days_to_exp, risk_free_rate = r
)
if price_range is not None:
ax.plot(price_range, pdf,
label=f"{exp.strftime('%Y-%m-%d')} ({days_to_exp}d)",
color=colors[idx % len(colors)], linewidth=2)
ax.fill_between(price_range, pdf, alpha=0.3, color=colors[idx % len(colors)])
# Add vertical line for current price
ax.axvline(x=current_price, linestyle='--', color='red', linewidth=2,
label=f'Current: ${current_price:.2f}')
ax.set_title(f'Implied Probability Distribution Functions - Overlay View looking at {TICKER}', fontsize=14, fontweight='bold')
ax.set_xlabel('Stock Price', fontsize=12)
ax.set_ylabel('Probability Density', fontsize=12)
ax.legend(loc='upper right', fontsize=9)
ax.grid(True, alpha=0.3)
plt.tight_layout()
return fig
def create_mode_and_quartiles(options_df, current_price):
"""
Create chart showing most probable value (mode) and confidence intervals
"""
expirations = sorted(options_df['expiration'].unique())
exp_labels = []
modes = []
q25 = []
q75 = []
q5 = []
q95 = []
days_list = []
for exp in expirations:
exp_data = options_df[options_df['expiration'] == exp]
days_to_exp = exp_data['days_to_exp'].iloc[0]
atm_option = exp_data.iloc[(exp_data['strike'] - current_price).abs().argsort()[:1]]
if len(atm_option) > 0:
iv = atm_option['impliedVolatility'].iloc[0]
T = days_to_exp / 365.0
# Mode of lognormal distribution
mode = current_price * np.exp(-iv**2 * T)
# Calculate quantiles using lognormal distribution
mu = np.log(current_price) + (0.05 - 0.5 * iv**2) * T
sigma = iv * np.sqrt(T)
q5_val = np.exp(mu + sigma * norm.ppf(0.05))
q25_val = np.exp(mu + sigma * norm.ppf(0.25))
q75_val = np.exp(mu + sigma * norm.ppf(0.75))
q95_val = np.exp(mu + sigma * norm.ppf(0.95))
exp_labels.append(f"{exp.strftime('%m/%d')}")
modes.append(mode)
q25.append(q25_val)
q75.append(q75_val)
q5.append(q5_val)
q95.append(q95_val)
days_list.append(days_to_exp)
fig, ax = plt.subplots(figsize=(12, 7))
x_pos = np.arange(len(exp_labels))
# 90% confidence interval
ax.fill_between(x_pos, q5, q95, alpha=0.3, color='lightblue', label='90% CI (5th-95th)')
# 50% confidence interval (IQR)
ax.fill_between(x_pos, q25, q75, alpha=0.5, color='blue', label='50% CI (25th-75th)')
# Mode (most probable value)
ax.plot(x_pos, modes, 'o-', color='darkgreen', linewidth=3,
markersize=8, label='Most Probable (Mode)')
# Current price line
ax.axhline(y=current_price, linestyle='--', color='red', linewidth=2,
label=f'Current: ${current_price:.2f}')
ax.set_title(f'Most Probable Value and Confidence Intervals by Expiration looking at {TICKER}',
fontsize=14, fontweight='bold')
ax.set_xlabel('Expiration Date', fontsize=12)
ax.set_ylabel('Stock Price', fontsize=12)
ax.set_xticks(x_pos)
ax.set_xticklabels(exp_labels, rotation=45, ha='right')
ax.legend(loc='best', fontsize=10)
ax.grid(True, alpha=0.3, axis='y')
plt.tight_layout()
return fig# Set ticker symbol
TICKER = "SPY" # Change this to any ticker
interest = yf.Ticker("^IRX")
r = interest.history(period="1d")["Close"].iloc[-1]/100 # assume risk-free rate ~5% (you could pull ^IRX)
print(f"Fetching option chain data for {TICKER}...")
options_df, current_price = get_option_chain_data(TICKER)
print(f"Current price: ${current_price:.2f}")
print(f"Total options loaded: {len(options_df)}")
print(f"Expirations available: {len(options_df['expiration'].unique())}")Fetching option chain data for SPY...
Current price: $663.68
Total options loaded: 7934
Expirations available: 30# Create visualizations
print("\nGenerating 3D IPDF surface plot...")
fig_3d = create_3d_ipdf_surface(options_df, current_price)
fig_3d.show()
Generating 3D IPDF surface plot...print("Generating IPDF overlay plot...")
fig_overlay = create_ipdf_overlay(options_df, current_price)
fig_overlay.show()Generating IPDF overlay plot...C:\Users\pietr\AppData\Local\Temp\ipykernel_18156\3480485104.py:3: UserWarning:
FigureCanvasAgg is non-interactive, and thus cannot be shown
print("Generating mode and quartiles plot...")
fig_quartiles = create_mode_and_quartiles(options_df, current_price)
fig_quartiles.show()Generating mode and quartiles plot...C:\Users\pietr\AppData\Local\Temp\ipykernel_18156\3409887583.py:192: RuntimeWarning:
invalid value encountered in sqrt
C:\Users\pietr\AppData\Local\Temp\ipykernel_18156\495272846.py:3: UserWarning:
FigureCanvasAgg is non-interactive, and thus cannot be shown
